Inverter-driven motors are widely used in many industrial applications. The insulation on the stator windings of these motors is subject to deterioration due to high transient voltages, inverter cable reflections, mechanical stress, high temperatures, and aging. As a result, the insulation can fail, resulting in inter-turn winding faults (shorts from one part of a winding to another part of the winding). These inter-turn faults reduce efficiency and may cause excess heating, which may eventually cascade to irreversible damage such as melted conductors and motor failure. Given the cost of an interruption of critical systems being driven by these motors, it is important to detect inter-turn winding faults before irreversible damage occurs. Typically, detection of inter-turn winding faults involves monitoring of the motor currents and voltages and using a processor to analyze any asymmetries of currents through the motor, or to analyze differences in motor currents between a healthy motor and one with winding faults. In some embodiments, the inverter switches are used to inject voltages at a frequency higher than the normal switching frequency, and the resulting currents are analyzed by a processor. Analysis of motor currents requires significant processor resources, and using the inverter's switches to inject higher frequency voltages can result in audible noise. There is an ongoing need for improved detection of motor faults.